Enterprise management-oriented enterprise operation analysis and supervision method and system

Abstract

The invention relates to the technical field of operation supervision, in particular to an enterprise management-oriented enterprise operation analysis supervision method and system, and the method comprises the steps: obtaining food propaganda data and quality inspection data uploaded by a manufacturer in food sales software; performing feature extraction on the propaganda data and the quality inspection data, and obtaining propaganda feature data and quality inspection feature data containing feature attraction levels according to attraction degrees of features to users; analyzing the matching degree of the quality inspection feature data and the propaganda feature data to obtain a feature matching result; and determining rectification measures of the manufacturer according to a matching result. According to the method, food supervision measures in multiple links are adopted for food data of manufacturers in the food sales software, so that the manufacturers possibly causing food safety events can be supervised, and the manufacturers can rectify and construct a good food sales environment; and the problem that a supervision department lacks supervision means in the aspect of food safety is solved.

Description

Technical Field

[0001] The present invention relates to the field of operation supervision technology, and in particular to an enterprise operation analysis and supervision method and system for enterprise management. Background Art

[0002] At present, in terms of food supervision, in various food sales software, manufacturers will upload various pictures, texts and other advertising information to attract consumers' attention; and the content promoted by some manufacturers is mostly inconsistent with the actual situation of the food, resulting in poor experience for consumers after purchase; and, often after a food safety incident occurs, the regulatory authorities will conduct targeted inspections on the manufacturers involved in the food safety incident, exposing the problem of the regulatory authorities' lack of prior regulatory measures for food safety.

[0003] Therefore, the present invention provides an enterprise operation analysis and supervision method and system for enterprise management to solve the above problems. Summary of the Invention

[0004] In view of the above situation, in order to overcome the defects of the existing technology, the present invention provides an enterprise operation analysis and supervision method and system for enterprise management to solve the problem of the lack of prior supervision measures for food safety by the above-mentioned regulatory departments.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] In the first aspect, the present invention provides an enterprise operation analysis and supervision method for enterprise management, comprising: after authorization, connecting to food sales software to obtain food promotion data and quality inspection data uploaded by manufacturers in the food sales software; based on the name of the food, extracting features from the promotion data and quality inspection data, and according to the degree of attraction of the features to users, obtaining promotion feature data and quality inspection feature data containing feature attraction levels; performing matching analysis on the quality inspection feature data and the promotion feature data to obtain feature matching results; when there is one mismatched feature in the first-level features, or two mismatched features in the second-level features, sending modification information to the manufacturer; when the uploaded information is qualified and the food is sold, obtaining consumer comment data; preprocessing and clustering analysis on the comment data to obtain clustering information including cluster type and cluster data; obtaining cluster data that is the same as the preset cluster type in the cluster information; using a hierarchical analysis model to perform supervision analysis on the cluster data to obtain supervision analysis results; obtaining the abnormal level in the supervision analysis results; and issuing rectification information to food manufacturers whose abnormal levels meet the abnormal preset conditions.

[0007] Preferably, the feature extraction of the promotional data is performed based on the food name, and promotional feature data including the feature attraction level is obtained according to the degree of attraction of the feature to the user, including: obtaining the promotional video data, promotional picture data and / or promotional text data in the promotional data, using a preset 3D convolutional neural network model to extract the spatiotemporal features of the promotional video data, using a preset convolutional neural network model to extract the image features of the promotional picture data, and using a preset language processing model to extract the text features of the promotional text data; mapping the spatiotemporal features, image features and text features to the same vector space; using an attention mechanism or a weighted fusion method to fuse features of different modalities to obtain a first fused feature; determining the attraction level of each first fused feature according to the degree of attraction of the first fused feature to the user, and obtaining the promotional feature data.

[0008] Preferably, the feature extraction of the quality inspection data is performed based on the food name, and quality inspection feature data including the feature attraction level is obtained according to the degree of attraction of the feature to the user, including: using a text extraction model to extract text from the quality inspection data to obtain quality inspection text data; using a preset language processing model to perform feature extraction on the food name and the quality inspection text data to obtain semantic features and quality inspection text features, fusing the semantic features and the quality inspection text features to obtain a second fused feature; determining the attraction level of each second fused feature according to the degree of attraction of the second fused feature to the user to obtain promotional feature data.

[0009] Preferably, the matching analysis of the quality inspection feature data and the publicity feature data to obtain the feature matching result includes: aligning the quality inspection feature data and the publicity feature data in descending order of attractiveness level, and calculating the similarity of each quality inspection feature and the publicity feature using cosine similarity; mapping the cosine similarity to a preset interval to obtain the sub-matching degree between the quality inspection feature and the corresponding publicity feature in each level; and integrating and analyzing the sub-matching degrees of each level to obtain the feature matching result.

[0010] Preferably, the preprocessing and cluster analysis of the comment data to obtain clustering information including clustering types and clustering data includes: performing data cleaning on the comment data to obtain first processed data; and processing the first processed data using the K-Means clustering algorithm to obtain clustering information including clustering types and clustering data.

[0011] Preferably, the use of a hierarchical analysis model to perform regulatory analysis on cluster data to obtain regulatory analysis results includes: the hierarchical analysis model includes a factor layer, an indicator layer, an abnormality analysis layer and a target layer; the factor layer is used to perform indicator classification analysis on the cluster data to determine the factor data of each indicator; the indicator layer is used to analyze the price change index, food safety index, and complaint index of the food through the factor data; the abnormality analysis layer is used to analyze the abnormal information of the food according to the factor data of the high-risk level corresponding to the indicator score to obtain the abnormality analysis result; the target layer is used to analyze the evaluation results of each indicator and the abnormality analysis results to obtain the regulatory analysis result.

[0012] Preferably, the factor layer is used to perform index classification analysis on cluster data and determine factor data of each index, including: using Apriori algorithm to perform association analysis on cluster type and each index, determining the cluster type associated with each index, and obtaining factor data of each index.

[0013] Preferably, the abnormal analysis layer is used to analyze the abnormal information of the food according to the factor data of the high-risk level corresponding to the indicator score to obtain the abnormal analysis result, including: determining the abnormal information and abnormal information type in the factor data, and counting the number of occurrences, forwarding frequency, number of comments and comment frequency of the abnormal information within a preset time; performing weighted average analysis on the number of occurrences, forwarding frequency, number of comments and comment frequency to obtain the abnormal information change trend index; determining the abnormal impact score according to the forwarding frequency, comment frequency and abnormal information change trend index; determining the abnormal level according to the abnormal impact score and the abnormal information type based on the abnormal level comparison table; and integrating the abnormal level, abnormal information change trend index and abnormal information to obtain the abnormal analysis result.

[0014] Preferably, after obtaining the quality inspection feature data, a method for enterprise operation analysis and supervision for enterprise management also includes: obtaining the quality inspection record data of the quality inspection department based on the quality inspection department information in the quality inspection feature data; searching from the quality inspection record data whether there is quality inspection data of the manufacturer; and when it does not exist, sending modification information to the manufacturer.

[0015] In a second aspect, the present invention provides an enterprise operation analysis and supervision system for enterprise management, comprising a memory, a processor, and a computer program stored on the memory, wherein the processor executes the computer program to implement the enterprise operation analysis and supervision method described in any one of the above technical solutions.

[0016] The beneficial effects of the present invention are:

[0017] 1. This invention implements multiple food regulatory measures for manufacturers' food data within food sales software, helping to monitor manufacturers that may cause food safety incidents, prompting them to make corrections and fostering a positive food sales environment. Furthermore, by analyzing and processing manufacturer review data on food sold, this helps to promptly identify food safety issues and notify manufacturers to rectify them. Furthermore, this invention addresses the problem of regulatory authorities' lack of pre-emptive monitoring measures for food safety incidents.

[0018] 2. By analyzing and processing the review data, the present invention can promptly discover the behaviors of manufacturers that disrupt the food sales market during the food sales process, and then can promptly make targeted treatment measures, which helps to reduce the adverse effects caused by the manufacturers' disruptive behaviors in the food sales market.

[0019] 3. The present invention can verify the authenticity of the quality inspection data uploaded by the manufacturer to prevent the food of the manufacturer with falsified quality inspection from flowing into the hands of consumers, which helps to reduce the risk of food safety and increases the means of supervision before safety incidents occur. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a schematic flow chart of an enterprise operation analysis and supervision method for enterprise management according to the present invention;

[0021] Figure 2 This is a schematic diagram of an enterprise operation analysis and supervision system for enterprise management according to the present invention. DETAILED DESCRIPTION

[0022] The following will refer to the attached Figure 1 To the attached Figure 2 The embodiments of the present invention are described in detail. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0023] A method for analyzing and supervising enterprise operations for enterprise management, as shown in the attached Figure 1 As shown, the following steps are included:

[0024] Step S11: After authorization by the company of the food sales software, connect to the background database of the food sales software to obtain the food promotion data and quality inspection data uploaded by the manufacturer in the food sales software.

[0025] Among them, promotional data includes picture data, video data or text data submitted by manufacturers regarding food packaging, efficacy, suitable population, expiration date, etc.; quality inspection data includes food quality inspection ingredient data, efficacy data, adverse reactions and other certification certificate data related to product quality.

[0026] Step S12: Based on the food name, feature extraction is performed on the promotional data and quality inspection data, and according to the degree of attraction of the features to the user, promotional feature data and quality inspection feature data including feature attraction levels are obtained.

[0027] Step S13: Perform matching analysis on the quality inspection feature data and the promotional feature data to obtain feature matching results; when there is one mismatching feature in the first-level features, or two mismatching features in the second-level features, false advertising is identified and modification information is sent to the manufacturer.

[0028] Among them, the attraction levels include three levels, and the order of levels is that the first level is the largest, the second level is the second, and the third level is the smallest.

[0029] Step S14: When the uploaded information is qualified and the food is sold, obtain the consumer's comment data; pre-process and cluster analyze the comment data to obtain cluster information including cluster type and cluster data; obtain cluster data of the same cluster type as the preset cluster type in the cluster information.

[0030] Step S15: Perform supervision analysis on the cluster data using a hierarchical analysis model to obtain supervision analysis results; obtain abnormality levels in the supervision analysis results; and issue rectification information to food manufacturers whose abnormality levels meet abnormality preset conditions.

[0031] Through the aforementioned multiple food supervision links, it is possible to monitor manufacturers that may cause food safety incidents, prompting these companies to make corrections and fostering a positive food sales environment. Furthermore, by analyzing and processing review data on manufacturers' food sales, it is possible to promptly identify food safety issues and notify manufacturers to rectify them. Furthermore, this invention addresses the lack of pre-event monitoring tools for food safety incidents, facilitating a shift from post-event supervision to pre-event warning, thereby reducing the probability of food safety incidents.

[0032] In one embodiment of the present invention, the feature extraction of promotional data is performed based on the food name, and promotional feature data including the feature attraction level is obtained according to the degree of attraction of the feature to the user, including: obtaining promotional video data, promotional picture data and / or promotional text data in the promotional data, using a preset 3D convolutional neural network model to extract the spatiotemporal features of the promotional video data, using a preset convolutional neural network model to extract the image features of the promotional picture data, and using a preset language processing model to extract the text features of the promotional text data; mapping the spatiotemporal features, image features and text features to the same vector space; using an attention mechanism or a weighted fusion method to fuse features of different modalities to obtain a first fused feature; determining the attraction level of each first fused feature according to the degree of attraction of the first fused feature to the user, to obtain the promotional feature data.

[0033] Specifically, based on a large amount of video data, image data, and text data of different foods, a preset 3D convolutional neural network model is constructed using a 3D convolutional neural network algorithm, a preset convolutional neural network model is constructed using ResNet, and a preset language processing model is constructed using BERT. After obtaining the features of the different modalities output by each model, the features of the different modalities are mapped to the same vector space, and a shared projection layer or Transformer is used to achieve cross-modal mapping. The features of the different modalities are then fused using an attention mechanism or a weighted fusion method to obtain the first fused features of the promotional video data. Based on the degree of appeal of each first fused feature to the user, the different first fused features are divided into different attractiveness levels to obtain promotional feature data.

[0034] The evaluation indicators for attractiveness include one or more of taste, efficacy, weight or quality, and additive content, and can be adjusted based on the type and function of the food. Furthermore, the order of attractiveness levels for each evaluation indicator is adjusted based on the name of the food. Based on a database of matching relationships between features and evaluation indicators, the evaluation indicator with the highest matching degree for each fused feature is determined, and the level of the evaluation indicator is used as the attractiveness level for that fused feature, thereby generating promotional feature data containing fused features of different attractiveness levels.

[0035] Through the setting method of this embodiment, the present invention can use a multimodal fusion method to perform feature extraction and feature classification on the food promotional data uploaded by the manufacturer, and obtain promotional feature data containing first fusion features of different attractiveness levels, which is helpful for subsequent matching analysis of the promotional feature data and quality inspection feature data to obtain accurate feature matching results.

[0036] In one embodiment of the present invention, the feature extraction of quality inspection data is performed based on the food name, and quality inspection feature data including the feature attraction level is obtained according to the degree of attraction of the feature to the user, including: using a text extraction model to extract text from the quality inspection data to obtain quality inspection text data; using a preset language processing model to perform feature extraction on the food name and the quality inspection text data to obtain semantic features and quality inspection text features, fusing the semantic features and the quality inspection text features to obtain a second fused feature; determining the attraction level of each second fused feature according to the degree of attraction of the second fused feature to the user to obtain promotional feature data.

[0037] Specifically, the quality inspection data is in the form of images. Grayscaling, global thresholding, median filtering, or Gaussian filtering are used to preprocess the data to enhance the text recognition within the data. EAST is used to detect text regions within the preprocessed data. A CRNN (Convolutional Recurrent Neural Network) is used to recognize the text and obtain character sequences. Duplicate or irrelevant characters are removed, and recognition errors are corrected using a dictionary to obtain quality inspection text data. A Transformer is used to fuse semantic features with quality inspection text features to obtain secondary fused features. Based on the degree of user appeal of each secondary fused feature, the different secondary fused features are classified into different attractiveness levels to obtain quality inspection feature data.

[0038] The evaluation indicators for attractiveness include one or more of taste, efficacy, weight or quality, and additive content, and can be adjusted based on the type and function of the food. Furthermore, the order of attractiveness ranking of each evaluation indicator is adjusted based on the name of the food. Based on a database of matching relationships between features and evaluation indicators, the evaluation indicator with the highest matching degree for each second fused feature is determined, and the level of the evaluation indicator is used as the attractiveness level of the second fused feature, thereby generating quality inspection feature data containing second fused features of different attractiveness levels.

[0039] Through the setting method of this embodiment, the present invention can use a multimodal fusion method to perform feature extraction and feature classification on the food names and quality inspection text data uploaded by the manufacturer after using a text extraction model to extract the quality inspection text data in the quality inspection data, and obtain quality inspection feature data containing second fusion features of different attractiveness levels, which is helpful for subsequent matching analysis of the promotional feature data and the quality inspection feature data to obtain accurate feature matching results.

[0040] In one embodiment of the present invention, the matching analysis of the quality inspection feature data and the publicity feature data to obtain the feature matching result includes: aligning the quality inspection feature data and the publicity feature data in descending order of attractiveness level, and calculating the similarity of each quality inspection feature and the publicity feature using cosine similarity; mapping the cosine similarity to a preset interval to obtain the sub-matching degree of the quality inspection feature and the corresponding publicity feature in each level; and integrating and analyzing the sub-matching degree of each level to obtain the feature matching result.

[0041] The preset interval is 0-1. The integration process includes: integrating the first-level quality inspection features and the promotional features for a first time with a sub-matching degree of zero; integrating the second-level quality inspection features and the promotional features for a second time with a sub-matching degree of zero; and integrating the third-level quality inspection features and the promotional features for a third time with a sub-matching degree of zero.

[0042] Furthermore, when a quality inspection feature exists but the corresponding publicity feature does not exist, the corresponding sub-matching degree is adjusted to 1 in order to avoid statistical errors in the regulatory data, which may lead to reduced accuracy.

[0043] Through the setting method of this embodiment, the present invention can reasonably perform matching analysis on the promotional feature data and the quality inspection feature data, and obtain reasonable and practical feature matching results, so as to determine through subsequent analysis whether the manufacturer is allowed to sell food on the food sales software.

[0044] In one embodiment of the present invention, the preprocessing and cluster analysis of comment data to obtain clustering information including clustering types and clustering data includes: performing data cleaning on the comment data to obtain first processed data; and processing the first processed data using a K-Means clustering algorithm to obtain clustering information including clustering types and clustering data.

[0045] Specifically, the comment data is first cleaned, including removing noise, processing missing values, removing stop words, and text normalization, to obtain first processed data; then, a text vectorization method such as Word2Vec or BERT is used to convert the first processed data into a dense vector; the dense vector is normalized, and then the clustering step in the K-Means clustering algorithm is used to process the normalized dense vector to obtain the cluster label to which each data point belongs, the center vector of each cluster, and the statistical information of the data points within the cluster; wherein, the cluster label is used as the clustering type; the center vector of each cluster and the statistical information of the data points within the cluster are used as the clustering information.

[0046] In one embodiment of the present invention, the use of a hierarchical analysis model to perform regulatory analysis on cluster data to obtain regulatory analysis results includes: the hierarchical analysis model includes a factor layer, an indicator layer, an anomaly analysis layer and a target layer; the factor layer is used to perform indicator classification analysis on the cluster data to determine the factor data of each indicator; the indicator layer is used to analyze the price change index, food safety index, and complaint index of the food through the factor data; the anomaly analysis layer is used to analyze the abnormal information of the food according to the factor data of the high-risk level corresponding to the indicator score to obtain the anomaly analysis result; the target layer is used to analyze the evaluation results of each indicator and the anomaly analysis result to obtain the regulatory analysis result.

[0047] By using a hierarchical analysis model to analyze the clustered data derived from the review data in terms of price, food safety, and complaints, a comprehensive regulatory analysis of the food product is obtained. Subsequent steps are then used to process any abnormal information in the regulatory analysis results and send them to the manufacturer for rectification. Furthermore, by analyzing and processing the review data, the present invention can promptly identify any disruptive behavior by manufacturers during the food sales process, enabling timely and targeted action to mitigate the adverse effects of such behavior.

[0048] The processing of the factor layer includes: using the Apriori algorithm to perform association analysis on the cluster type and each indicator, determining the cluster type associated with each indicator, and obtaining the factor data of each indicator.

[0049] Among them, the types of factor data of price change indicators include: raw material costs, transportation costs, market demand, publicity costs, etc.

[0050] The data types of food safety indicators include heavy metal name and content, food additive name and content, illegal additives, storage conditions, shelf life, packaging materials, etc.

[0051] The factors of complaint indicators include taste and quality, weight, packaging issues, distribution issues, hygiene issues, expired food, food poisoning, labeling issues, publicity issues, supply chain issues, price issues, etc.

[0052] The analysis process of the indicator layer includes: when determining the score of each indicator based on the factor data of each indicator, calculating the indicator score of each indicator according to the statistical method preset for each indicator.

[0053] The calculation process for the price change index score includes: preprocessing the factor data of the price change index and forming a time series of food prices in chronological order; analyzing the time series data and statistically analyzing the influencing data that affect the price change index, including: calculating the price volatility based on the maximum, minimum, and mean values in the time series; calculating the price change rate based on price changes within a preset time period; calculating the price index based on the price of the food when it was first sold and the current price; calculating the raw material cost change rate based on the price changes of raw materials within a preset time period; and finally, calculating and formatting each influencing data using the weighted average method to obtain the price change index score. The corresponding relationship between the indicator score and the risk level is set in the corresponding relationship between the indicator score of the price change index and the risk level, where low risk: the indicator score is less than or equal to 20 points; medium risk: the indicator score is greater than 20 points and less than or equal to 50 points; and high risk: the indicator score is greater than 50 points.

[0054] The calculation process for food safety indicator scores involves weighting the scores for chemical, biological, and physical properties to determine the food safety indicator score. Each score is determined based on the actual test value and the standard limit for each aspect. A corresponding relationship is established between the food safety indicator score and the risk level: low risk: the indicator score is less than or equal to 100 points; medium risk: the indicator score is greater than 100 points and less than or equal to 200 points; and high risk: the indicator score is greater than 200 points.

[0055] Similarly, the analysis process of the indicator score of the complaint index is as follows: determine the first impact score based on the ratio of the number of complaints to the overall number of comments, with the maximum value of the first impact score being 100 and the minimum value being 0; and perform statistics on the types and amounts of complaints; analyze complaint data of the same type whose amount of complaints exceeds the preset value, and determine the second impact score based on the relationship between the complaint type and amount and the preset corresponding table; and then use the weighted average method to perform weighted average processing on the first impact score and multiple second impact scores to obtain the indicator score of the complaint index.

[0056] In the correspondence between indicator scores and risk levels, a correspondence between indicator scores and risk levels of food safety indicators is set, among which low risk: indicator score is less than or equal to 20 points; medium risk: indicator score is greater than 20 points and less than or equal to 50 points; high risk: indicator score is greater than 50 points.

[0057] The analysis process of the anomaly analysis layer includes: determining the abnormal information and abnormal information type in the factor data, counting the number of occurrences, forwarding frequency, number of comments and comment frequency of abnormal information within a preset time; performing weighted average analysis on the number of occurrences, forwarding frequency, number of comments and comment frequency to obtain the abnormal information change trend index; determining the abnormal impact score based on the forwarding frequency, comment frequency and abnormal information change trend index, and determining the abnormal level based on the abnormal level comparison table according to the abnormal impact score and abnormal information type; and integrating the abnormal level, abnormal information change trend index and abnormal information to obtain the abnormal analysis results.

[0058] Among them, the number of occurrences refers to the number of times the abnormal information itself appears in the comment data; the forwarding frequency refers to the frequency of forwarding to other users or group chats in various food sales software; the number of comments refers to the total number of all data commenting on the abnormal information; the comment frequency refers to the number of comments on the abnormal information per unit time.

[0059] Specifically, based on the analysis of the indicator layer, the abnormal information and abnormal information type in the factor data are determined, and the number of occurrences, forwarding frequency, number of comments and comment frequency of abnormal information within the preset time are counted; the number of occurrences, forwarding frequency, number of comments and comment frequency are weighted averaged to obtain the abnormal information change trend index; and the standardized forwarding frequency, comment frequency and abnormal information change trend index are processed using the weighted average method to determine the abnormal impact score; wherein the weight is a numerical value determined based on expert experience, and different forwarding frequencies, comment frequencies and abnormal information change trend indices correspond to different weight values.

[0060] Based on the anomaly level comparison table, the anomaly level is determined according to the anomaly impact score and the anomaly information type; different anomaly information types are set with different anomaly level lists; finally, the anomaly level, anomaly information change trend index and anomaly information are integrated to obtain the anomaly analysis results.

[0061] The pre-condition for an exception is that the exception level is high risk. When high-risk food information is found, the abnormal information corresponding to the high-risk level is integrated with the food manufacturer information, and rectification information is sent to the manufacturer, thereby strengthening supervision.

[0062] Through the analysis method of this embodiment, the present invention can use a hierarchical analysis model to process cluster data at the factor level, indicator level, anomaly analysis level, and target level, thereby obtaining comprehensive regulatory analysis results for the food product. Furthermore, by analyzing and processing review data, the present invention can promptly identify any disruptive behavior by manufacturers during the food sales process, enabling timely and targeted action to mitigate the adverse effects of such behavior.

[0063] In one embodiment of the present invention, after obtaining the quality inspection feature data, it also includes: obtaining the quality inspection record data of the quality inspection department based on the quality inspection department information in the quality inspection feature data; searching from the quality inspection record data whether there is quality inspection data of the manufacturer; when it does not exist, sending modification information to the manufacturer.

[0064] Specifically, based on the quality inspection information in the quality inspection feature data, connect to the quality inspection department platform and access the quality inspection record data on the platform; then search the quality inspection record data to see whether there is quality inspection data of the manufacturer corresponding to the date in the quality inspection feature data. If it cannot be found, it is determined that the quality inspection data uploaded by the manufacturer is incorrect and is false advertising. A modification message is sent to the manufacturer to ask the manufacturer to upload the correct quality inspection data.

[0065] Through the setting method of this embodiment, the present invention can verify the authenticity of the quality inspection data uploaded by the manufacturer to prevent the food of the manufacturer with falsified quality inspection from flowing into the hands of consumers, which helps to reduce the risk of food safety, and increases the early warning means before the occurrence of safety incidents through the above method.

[0066] In one embodiment of the present invention, Figure 2 As shown, it includes a memory, a processor and a computer program stored in the memory, and the processor executes the computer program to implement the enterprise operation analysis and supervision method described in any one of the above embodiments.

[0067] By combining the enterprise operation analysis and supervision methods described in the above embodiments, the present invention can supervise food sales at multiple levels, helping to supervise manufacturers that may cause food safety incidents, allowing these companies to make rectifications and building a favorable food sales environment. Furthermore, by analyzing and processing manufacturer review data on food sold, it helps to promptly identify food safety issues and notify manufacturers to rectify them. Furthermore, the present invention addresses the problem of regulatory authorities' lack of pre-event supervision measures for food safety incidents, facilitating a shift from post-event supervision to pre-event warnings, thereby helping to reduce the probability of food safety incidents.

[0068] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard processors (ASSPs), systems on a chip (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0069] It should be noted that, in the description of the present invention, the terms "first", "second" and "third" are used for descriptive purposes only and should not be understood as indicating or implying relative importance.

[0070] The program code for implementing the method of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device so that when the program code is executed by the processor or controller, the functions / operations specified in the flow chart and / or block diagram are implemented. The program code can be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0071] In the context of the present disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in conjunction with an instruction execution system, device or equipment. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium can include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0072] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0073] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0074] A computer system may include a client and a server. The client and server are generally remote from each other and typically interact through a communication network. The client-server relationship arises through computer programs running on the respective computers and having a client-server relationship with each other. The server may be a cloud server, a server in a distributed system, or a server integrated with a blockchain.

[0075] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

Claims

1. A method for analyzing and supervising enterprise operations for enterprise management, characterized in that: include: After authorization, connect to the food sales software to obtain food promotion data and quality inspection data uploaded by manufacturers in the food sales software; Based on the food name, feature extraction is performed on the promotional data and quality inspection data, and according to the attractiveness of the features to users, promotional feature data and quality inspection feature data containing feature attractiveness levels are obtained; Perform matching analysis on the quality inspection feature data and the promotional feature data to obtain feature matching results; when there is one mismatching feature in the first-level features, or two mismatching features in the second-level features, send modification information to the manufacturer; Once the uploaded information is qualified and the food is sold, obtain consumer review data; Preprocessing and cluster analysis are performed on the comment data to obtain cluster information including cluster type and cluster data; Obtain cluster data of the same type as the preset cluster in the cluster information; A hierarchical analysis model is used to conduct supervisory analysis on cluster data to obtain supervisory analysis results; the abnormality level in the supervisory analysis results is obtained; and rectification information is issued to food manufacturers whose abnormality levels meet the abnormality preset conditions.

2. The enterprise operation analysis and supervision method according to claim 1, characterized in that: The method of extracting features from promotional data based on food names and obtaining promotional feature data including feature attraction levels according to the degree of attraction of the features to users includes: obtaining promotional video data, promotional picture data and / or promotional text data from the promotional data, extracting spatiotemporal features of the promotional video data using a preset 3D convolutional neural network model, extracting image features of the promotional picture data using a preset convolutional neural network model, and extracting text features of the promotional text data using a preset language processing model; mapping the spatiotemporal features, image features and text features to the same vector space; fusing features of different modalities using an attention mechanism or a weighted fusion method to obtain a first fused feature; and determining the attraction level of each first fused feature according to the degree of attraction of the first fused feature to the user to obtain the promotional feature data.

3. The enterprise operation analysis and supervision method according to claim 1, characterized in that: The method of extracting features from quality inspection data based on food names and obtaining quality inspection feature data including feature attraction levels according to the degree of attraction of the features to users includes: extracting text from the quality inspection data using a text extraction model to obtain quality inspection text data; extracting features from food names and quality inspection text data using a preset language processing model to obtain semantic features and quality inspection text features, fusing the semantic features and quality inspection text features to obtain second fused features; and determining the attraction level of each second fused feature according to the degree of attraction of the second fused features to users to obtain promotional feature data.

4. The enterprise operation analysis and supervision method according to claim 1, characterized in that: The matching analysis of the quality inspection feature data and the publicity feature data to obtain the feature matching result includes: aligning the quality inspection feature data and the publicity feature data in descending order of attraction level, and calculating the similarity between each quality inspection feature and the publicity feature using cosine similarity; mapping the cosine similarity to a preset interval to obtain the sub-matching degree between the quality inspection feature and the corresponding publicity feature in each level; and integrating and analyzing the sub-matching degrees of each level to obtain the feature matching result.

5. The enterprise operation analysis and supervision method according to claim 1, characterized in that: The preprocessing and cluster analysis of the comment data to obtain clustering information including clustering types and clustering data includes: data cleaning of the comment data to obtain first processed data; and processing the first processed data using the K-Means clustering algorithm to obtain clustering information including clustering types and clustering data.

6. The enterprise operation analysis and supervision method according to claim 1, characterized in that: The described method of using a hierarchical analysis model to perform regulatory analysis on cluster data to obtain regulatory analysis results includes: the hierarchical analysis model includes a factor layer, an indicator layer, an abnormality analysis layer and a target layer; the factor layer is used to perform indicator classification analysis on the cluster data and determine the factor data of each indicator; the indicator layer is used to analyze the price change index, food safety index, and complaint index of the food through the factor data; the abnormality analysis layer is used to analyze the abnormal information of the food according to the factor data of the high-risk level corresponding to the indicator score to obtain the abnormality analysis result; the target layer is used to analyze the evaluation results of each indicator and the abnormality analysis results to obtain the regulatory analysis result.

7. The enterprise operation analysis and supervision method according to claim 6, characterized in that: The factor layer is used to perform index classification analysis on cluster data and determine the factor data of each index, including: using Apriori algorithm to perform association analysis on cluster types and each index, determining the cluster type associated with each index, and obtaining the factor data of each index.

8. The enterprise operation analysis and supervision method according to claim 6, characterized in that: The abnormality analysis layer is used to analyze the abnormal information of the food according to the factor data of the high-risk level corresponding to the indicator score to obtain the abnormality analysis results, including: determining the abnormal information and abnormal information type in the factor data, and counting the number of occurrences, forwarding frequency, number of comments and comment frequency of the abnormal information within a preset time; performing weighted average analysis on the number of occurrences, forwarding frequency, number of comments and comment frequency to obtain the abnormal information change trend index; determining the abnormal impact score according to the forwarding frequency, comment frequency and abnormal information change trend index; determining the abnormality level according to the abnormal impact score and the abnormal information type based on the abnormality level comparison table; and integrating the abnormality level, abnormal information change trend index and abnormal information to obtain the abnormality analysis results.

9. The enterprise operation analysis and supervision method according to claim 1, characterized in that: After obtaining the quality inspection feature data, it also includes: obtaining the quality inspection record data of the quality inspection department based on the quality inspection department information in the quality inspection feature data; searching from the quality inspection record data whether there is quality inspection data of the manufacturer; when it does not exist, sending modification information to the manufacturer.

10. An enterprise operation analysis and supervision system for enterprise management, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the enterprise operation analysis and supervision method according to any one of claims 1 to 9.

Images